™
™Compare context before numbers
Two projects can mention the same gene and still answer different questions. One may report a binding assay; another may report a clinical outcome. Their measurements cannot be treated as interchangeable simply because search terms match.
The engine supports project, experiment, cohort, sample, compound, treatment, target, pathway, signal, method, result, and outcome comparisons within authorized project records. Each comparison checks its context before reporting agreement or difference.
Show agreement, disagreement, and absence
A side-by-side view should expose shared context, deviations, contradictory findings, missing fields, and evidence that was not measured or not profiled. These are different states. An empty field must not become a zero or a negative result.
A GDC record and a cBioPortal record may trace to the same study or samples. Two API responses do not become two independent confirmations. The workspace records source-overlap relationships and keeps lineage visible in governed comparisons. It does not treat duplicated or overlapping evidence as independent confirmation.
What the current illustration means
The homepage comparison board illustrates dimensions a researcher should inspect. It is not a live scientific comparison, ranking, or validated result. A researcher must review the original studies and the rights attached to each source before sharing any conclusion.